480 CHEMICAL ENGINEERING The Holborn-Kurlbaum pyrometer must be carefully calibrated for the particular absorption glass in use. It should be kept in mind that the lamp needs to be recalibrated from time to time. The calibration methods of Kanolt are to be recommended for this purpose.1 In work involving direct pyrometric determinations it is usually not feasible to take the bending of the test cones as the softening point but the end point is considered to have been reached when a distinct rounding of the edges is observed or a bead is formed. At best the estimation of the so-called fusion or softening point is not so easy a matter as might appear, both owing to the difficulty of close observation and the fact that at very high temperatures volatilization becomes quite marked, manifesting itself in the apparent subliming or melting away of the test specimen, especially with magnesia or its compounds. The effect of the intense reduction almost unavoidable in the electric furnace also is very prominent so that the real melting point of the substance under test can not be found, due to the reduction of the iron oxide commonly present to metallic iron as well as its actual volatilization. Silica also is driven off in appreciable amounts and any alkalis present are lost completely. Where the character of the refractory permits it the fusion test should be made in the gas-fired furnace under neutral or at least not excessively reducing conditions. The electric furnace then remains for the testing of materials fusing above the melting point of platinum, 1,755°C. Permanent Expansion and Contraction.—Practically all refractories are subject to further changes in volume when heated for longer periods under the conditions of their use. These volume changes are called permanent in distinction from those brought about by thermal expansion and contraction. Clay, magnesite and alumina refractories tend continually to contract or shrink while the siliceous materials expand more or less. It is evident that excessive volume changes are exceedingly undesirable inasmuch as they tend to bring about a breaking down of the furnace structure, through the opening up of joints, and the formation of cracks, thus causing leakage and ultimate collapse of the furnace. For most purposes materials showing as small a contraction or expansion in use are most desirable. Excessive shrinkage may be due to insufficient firing or to lack of refractoriness. For most purposes the harder a refractory is fired in its manufacture the more satisfactory it is. The contraction of a material with temperature may be employed as a means of estimating its refractoriness as well as for determining the proper point to which it should be fired. This affords a method of study which is exceedingly valuable and in fact makes possible a classification of refractories and the discrimination between the various grades. In securing the data needed for this purpose it is only necessary to prepare a series of specimens from the unburned material of such size as will permit accurate measurement of volume and to fire them at a number of increasing temperatures. We might, for instance, mold and dry 12 briquettes, 4 by 3 by 3 cm., immersing them in petroleum until completely saturated and then determine their volume by any convenient method. After heating the specimens to a low temperature just sufficient to expel the petroleum they are placed in a suitable test kiln or furnace capable of good temperature control and provided with calibrated platinum-platinum rhodium thermocouples and a high-resistance millivoltmeter or, still better, a potentiometer. The temperature range must be adjusted to the kind of refractory being tested. If we are dealing with a fireclay material the range between 1,150° and 1,425°C. may be chosen. The heat is raised from above 1,000°C. at the rate of 30CC. 1 Bureau of Standards, Technologic Paper No. 10, see also p. 450, this book.